In February 2012, a dozen members of the Food and Drug Administration’s Neurological Device Panel spent a Friday on the stage of a Hilton ballroom hearing about cranial electrotherapy stimulation (CES) devices. Tracey Kirsch was in the audience while the panel was warned by FDA personnel no fewer than 25 times that CES—a weak, pulsed electrical current delivered across the head—causes seizures. Kirsch runs a company that produces AlphaStim, one such device that attaches to the ear lobes and may treat anxiety, depression, and insomnia by briefly inhibiting activity in the brain’s cortex.
The hearing in Washington would determine whether CES devices would be reclassified from a Class III device (meant for life-sustaining products like heart valves) to a Class II device (alongside powered wheelchairs and pregnancy tests). Having used her allotted 45 minutes in the morning to argue for reclassification, Kirsch could not rise to object that not a single seizure had been caused in the 30-year marketing history of the product. Despite their largest customer being Veteran’s Affairs, many of whose patients are routinely offered CES treatment in VA facilities, there seems to be a mistrust of the technology at the FDA, expressed through sometimes-arbitrary criteria that discard the majority of supporting evidence. Eventually a witness clarified that the one documented seizure had happened during a clinical trial’s five-day drug washout period, before the electrotherapy had even begun. Unlike CES, benzodiazepine withdrawal frequently causes seizures.
“There were 50 people from the FDA at the front, facing the panel, and they were high-fiving, texting each other, breaking into the panel discussions,” says Kirsch. “They were out to win.”
In the face of frustration from disenchanted business owners like Kirsch, the FDA has the unenviable task of protecting Americans from medical dangers and duds. It must marshal all of the data on safety and effectiveness of new interventions, from cough syrup to spinal implants. Balancing those twin factors can be a haphazard and subjective affair, and it does not stop when a license to bring a product to market is issued. As a result, each person who relies on a given medicine is a data point in the society-wide assessment of that medicine’s healing powers. In coming years, Americans will all have a role to play in the ongoing research on drugs we take every day.
Working incognito in Chicago’s stockyards, Upton Sinclair couldn’t have known what an upheaval his 1905 novelization would cause. The Jungle documented exploitation of meat packers, as well as unsanitary handling of meat. It would provoke an uproar that led to a series of laws, including the 1906 Pure Food and Drugs Act that minted the FDA as a protection agency.
In its current incarnation, the FDA oversees a vast array of products, including food, pharmaceuticals, and medical devices. Drugs may be advertised and companies may vie for market share, but FDA regulation attempts to ensure that companies cannot compete with one another by cutting corners that put the safety of their customers at risk. All potential side effects must be out in the open and deemed worth the risk. Neither can those companies profit from selling a substandard product that does not achieve the effects they claim it does.
Already historically a sluggish process, licensing seems to have slowed to a crawl in recent years. Take for example some hypothetical new device that is similar in form and function to one that is already established on the market. In 1976, the FDA would have taken 38 days to process the completed application for classification. By 2015, the average time was 172 days, on top of years of clinical research.
“Obtaining clinical data takes time, typically three or more years for even a small pilot study,” says Kelly Roman, of Fisher-Wallace Laboratories, another CES company.
Medical-device companies—unlike drug conglomerates—are predominantly small businesses, and the process of bringing their inventions to market can be prohibitively expensive. After a device is finally cleared, its design may become obsolete within a couple of years.
For drug licensing, part of the time burden comes from the FDA’s dueling mandates of safety and efficacy, each of which requires different kinds of tests and controls. Each stage of research emphasizes the two factors to different extents, and these priorities are reflected in the experimental design. The weighting of safety and efficacy priorities remains a subject of debate.
For example, a new vaccine introduces risk to a broad section of an otherwise healthy population, while making them less vulnerable to a specific disease. In this case, any side effects will weigh heavily in the benefit-to-risk ratio, because the benefit of immunity is small when spread across millions of people, most of whom likely would not have contracted the disease anyway.
“I think the FDA needs to get out of the efficacy business,” says Daniel Kirsch, Tracey’s husband and business partner. “They should just focus on safety.”
The psychiatrist Jason Worchel agrees: “This is almost an insult to those of us who are physicians. Of course we monitor our patients, and if it’s ineffective we would change [the prescription].”
In practice, the earliest efficacy studies are not even contemplated until safety studies have been done, first using animals and then human subjects. Mary Foulkes, the former director of the FDA’s Office of Biostatistics and Epidemiology, informally describes the process: “We try it on one human being. Well, that didn’t seem to kill them. Then we try it on another.”
The next step is a dose evaluation, to find the maximum tolerated dose. Sometimes, safety is the limit set by a subject saying, “Please, no more. I can’t take it.”
Kirsch’s idea is that if the efficacy question is set aside, American citizens’ right to safety might be more easily guarded, even as the regulation process is streamlined. Efficacy then could be sorted out in the medical journals and the market. But on closer inspection, this doesn’t hold up, because safety and efficacy are inextricably entwined. Not just a mitigating factor against safety fears, efficacy is in fact a component of safety, which is defined as “a judgment of the acceptability of risk in a specified situation.” In other words, a product is deemed safe if its risks are more than compensated by its effectiveness at combatting the health problem, especially if that health problem is severe. Still, there is a good argument to be made that safety is not sufficiently weighted against efficacy.
Foulkes likens the data-set on the effects of any given drug to Swiss cheese. As much as researchers try to shrink those holes in the excel sheet where data is missing, they must ultimately make assumptions about what would fill those holes if the data was there. When it comes to efficacy data, the rules for filling in those holes are well spelled out. In analyzing safety outcomes, though, late-stage clinical trials often take a less-methodical approach. The current system is particularly bad at measuring adverse reactions that emerge over the course of a trial, as opposed to anticipated reactions for which experimenters are on the lookout.
“With a new class of drug, the safety issues are largely unknown,” says Foulkes. “If we take the case of Ebola, we were flying blind.”
As a statistician, Foulkes would like to see more rigor all the way through the drug-approval process. “There are a lot of patient testimonials presented to the FDA advisory boards,” she says. “Those are just anecdotes, and they provide very little information compared with confirmatory randomized controlled trials.”
Drugs are necessarily licensed before long-term hazards come to light, so post-licensing surveillance continues indefinitely. Efficacy, too, is a continuously shifting and nuanced concept, and some drugs become less effective over time as people’s pathogens adjust to them. Without monitoring in the population long after licensing, we might never know it.
In the future, Foulkes is hopeful that electronic medical records will capture far more of the problems with drugs once they’re released into the world than FDA’s current reporting mechanism. The FDA’s MedWatch program receives thousands of reports of adverse reactions every day, but this only captures an estimated 30 percent of the serious events. There are no obligations for physicians to report symptoms, because unlike trial subjects, patients don’t agree to share information when they take a drug. The whole of society then becomes a Swiss cheese of data, with no guarantee that the patients in the holes are happy with their treatment.
Which brings us to civic duty. The bioethicist Ezekiel Emanuel argues that people are morally obligated to participate in clinical trials, in just the same way that people are obligated to donate blood if healthy and able. Perhaps that obligation to future generations goes further. In an interconnected world, citizens would do well to view their own health in the context of their nation’s health and, indeed, the globe’s.
As the world faces a crisis in antibiotics resistance as well as increasing danger of spillover diseases from other species, it will take everyone to uphold effective healthcare. Given the size and complexity of the shared health community, reporting symptoms is the only way forward.
Meanwhile, setting aside the human foibles that plague all of our systems, clinical research continues to advance our wellbeing. “If I were on my deathbed and the doctor said I might be eligible as a subject for a new treatment, I know enough about clinical trials—I would jump at that,” says Foulkes.
In turn-of-the-century America, that’s a lot more than Upton Sinclair could have said.